III. BACTERIOPHAGE DNA AND BACTERIAL DNA 143 



investigations is the thyniinelcss death phenomena discovered by Cohen 

 and Earner (1954). When thymine is abruptly removed from a mutant 

 strain of coli (15T ) wliich has a requirement for thymine, the cells 

 gradually lose their viability in an exponential fashion for over four 

 decades. In the case of the triple mutant T-A-U' thymineless death pro- 

 ceeds essentially as before when the cells are incubated in a medium 

 lacking T, A, and U except that there is a residual fraction (3% to 

 10%) of the cells which is immune to thymineless death. This fraction 

 cen be increased to 100% of the cells by a 90-minute preincubation in 

 a medium containing T, but lacking A and U. These authors make the 

 plausible interpretation that this preincubation period allows the cells 

 which have begun DNA synthesis to complete the templating of a 

 genome. Before another round of templating can begin, there are some 

 essential steps which require arginine and uracil, which were missing 

 from this medium. This suggests that thymineless death is the result of 

 attempted DNA synthesis which can take place in the absence of 

 thymine only at a very slow rate and with a high incidence of errors, 

 some of which are lethal. 



If it were true that only a completion of a round of replication were 

 permissible in the presence of thymine, and in the absence of arginine 

 and uracil, then the resulting molecules after being broken during ex- 

 traction should have hybrid density if 5-bromouracil were added in 

 place of thymine. This hypothesis would predict that no fully heavy 

 molecules would be found even on prolonged incubation with 5-bromo- 

 uracil. On banding the extracted DNA, two distinct, symmetrical bands 

 were found which corresponded to the density of normal DNA and to 

 DNA of hybrid density. Thus, segments of the expected form were 

 found. This conclusion is weakened somewhat by the control experi- 

 ments which indicated that it was difficult to form uniformly BU-Iabeled 

 DNA; a fact which is probably related to an inhibitoiy effect of 

 bromouracil (Maal0e, 1962). 



Taking these results, together with those of Forro, strongly suggests 

 that the bacterial chromosome is a single structure of fixed DNA con- 

 tent, and that the two independently labeled units that emerge from 

 the studies of Forro are united in equal proportions in such a way as to 

 produce hybrid material in the density gradient. Since the hybrid DNA 

 can be broken extensively by sonication and yet retain its hybrid density, 

 one is forced to admit that traverse breaks do not separate the subunits 

 (Rolfe, 1962). This is exactly what one would expect if the two inde- 

 pendent labeling units were the complementary polynucleotide chains 

 making up the duplex DNA molecule. On the other hand, one would 

 expect the same thing from a pair of duplex molecules united at frequent 



